def main(cluster_file, consensus_file):
with mgf.read(consensus_file) as reader:
for spectrum_dict in reader:
peptide_seq = spectrum_dict['params']['title']
precursor_mz = spectrum_dict['params']['pepmass'][0]
precursor_charge = spectrum_dict['params']['charge'][0]
cons_mz = spectrum_dict['m/z array']
cons_intensity = spectrum_dict['intensity array']
retention_time = float(spectrum_dict['params']['rtinseconds'])
break
cons_spec = sus.MsmsSpectrum(peptide_seq,
precursor_mz=precursor_mz,
precursor_charge=precursor_charge,
mz=cons_mz,
intensity=cons_intensity,
retention_time=retention_time,
peptide=peptide_seq)
with mgf.read(cluster_file) as reader:
for spectrum_dict in reader:
precursor_mz = spectrum_dict['params']['pepmass'][0]
precursor_charge = spectrum_dict['params']['charge'][0]
mz = spectrum_dict['m/z array']
intensity = spectrum_dict['intensity array']
retention_time = float(spectrum_dict['params']['rtinseconds'])
spectrum = sus.MsmsSpectrum(peptide_seq,
precursor_mz=precursor_mz,
precursor_charge=precursor_charge,
mz=mz,
intensity=intensity,
retention_time=retention_time,
peptide=peptide_seq)
# Process the MS/MS spectrum.
fragment_tol_mass = 10
fragment_tol_mode = 'ppm'
# fragment_tol_mass = .5
# fragment_tol_mode = 'Da'
spectrum = (spectrum.set_mz_range(
min_mz=100, max_mz=1400).remove_precursor_peak(
fragment_tol_mass, fragment_tol_mode).filter_intensity(
min_intensity=0.05, max_num_peaks=50).scale_intensity(
'root').annotate_peptide_fragments(fragment_tol_mass,
fragment_tol_mode,
ion_types='aby'))
# Generate theoretical spec
# Plot the MS/MS spectrum.
fig, ax = plt.subplots(figsize=(12, 6))
# sup.spectrum(spectrum, ax=ax)
sup.mirror(spectrum, tspec, ax=ax)
plt.show()
plt.close()
def plot_spectrum(identifier, precursor_mz, precursor_charge, mz, intensity,
retention_time, peptide):
# identifier = spectrum_dict['params']['title']
# precursor_mz = spectrum_dict['params']['pepmass'][0]
# precursor_charge = spectrum_dict['params']['charge'][0]
# mz = spectrum_dict['m/z array']
# intensity = spectrum_dict['intensity array']
# retention_time = float(spectrum_dict['params']['rtinseconds'])
# peptide = 'WNQLQAFWGTGK'
# Create the MS/MS spectrum.
spectrum = sus.MsmsSpectrum(identifier,
precursor_mz=precursor_mz,
precursor_charge=precursor_charge,
mz=mz,
intensity=intensity,
retention_time=retention_time,
peptide=peptide)
# Process the MS/MS spectrum.
# fragment_tol_mass = 10
# fragment_tol_mode = 'ppm'
fragment_tol_mass = .5
fragment_tol_mode = 'Da'
spectrum = (spectrum.set_mz_range(
min_mz=100, max_mz=1400).remove_precursor_peak(
fragment_tol_mass, fragment_tol_mode).filter_intensity(
min_intensity=0.05, max_num_peaks=50).scale_intensity(
'root').annotate_peptide_fragments(fragment_tol_mass,
fragment_tol_mode,
ion_types='aby'))
# Generate theoretical spec
ts = sus._get_theoretical_peptide_fragments(peptide)
tmz = [frag.calc_mz for frag in ts]
ti = [1.0 for _ in tmz]
tspec = sus.MsmsSpectrum(identifier,
precursor_mz=precursor_mz,
precursor_charge=precursor_charge,
mz=tmz,
intensity=ti,
retention_time=retention_time,
peptide=peptide)
# Plot the MS/MS spectrum.
fig, ax = plt.subplots(figsize=(12, 6))
# sup.spectrum(spectrum, ax=ax)
sup.mirror(spectrum, tspec, ax=ax)
plt.show()
plt.close()
def test_init_modification_str():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
with pytest.raises(ValueError):
spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity,
peptide='PEPTIDER', modifications={'bla': 42})
def test_init_peptide():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
with pytest.raises(ValueError):
spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity,
peptide='BJOUXZ')
def get_cluster_spectra(mgf_filename: str) -> Dict[str, sus.MsmsSpectrum]:
"""
Read all spectra from the given MGF file corresponding to a single cluster.
Parameters
----------
mgf_filename : str
The file name of the MGF file to be read.
Returns
-------
Dict[str, sus.MsmsSpectrum]
A dictionary with as keys the scan numbers and as values the
corresponding spectra.
"""
spectra = {}
for spectrum_dict in mgf.read(mgf_filename):
# TODO: Make sure the USI doesn't contain a peptide identification.
cluster, usi = spectrum_dict['params']['title'].split(';')
spectrum = sus.MsmsSpectrum(
usi,
spectrum_dict['params']['pepmass'][0],
spectrum_dict['params']['charge'][0],
spectrum_dict['m/z array'],
spectrum_dict['intensity array'],
retention_time=spectrum_dict['params']['rtinseconds'])
spectrum.cluster = cluster
if usi in spectra:
raise ValueError(f'Non-unique USI: {usi}')
spectra[usi] = spectrum
return spectra
def fraction_of_by(peptide_seq, precursor_mz, precursor_charge, mz, intensity):
if not parser.fast_valid(peptide_seq):
print("Invalid peptide sequence encountered", file=sys.stderr)
return 0.0
spec = sus.MsmsSpectrum(peptide_seq,
precursor_mz=precursor_mz,
precursor_charge=precursor_charge,
mz=mz,
intensity=intensity,
peptide=peptide_seq)
fragment_tol_mass = 50
fragment_tol_mode = 'ppm'
spectrum = (spectrum.set_mz_range(
min_mz=100, max_mz=1400).remove_precursor_peak(
fragment_tol_mass,
fragment_tol_mode).annotate_peptide_fragments(fragment_tol_mass,
fragment_tol_mode,
ion_types='by'))
current, by_current = 0., 0.
for ix in range(len(spectrum.intensity)):
current += spectrum.intensity[ix]
if spectrum.annotation[ix] != None:
by_current += spectrum.intensity[ix]
if current > 0.:
return by_current / current
else:
return 0.0
def _parse_spectrum(spectrum_dict: Dict) -> sus.MsmsSpectrum:
"""
Parse the Pyteomics cluster dict.
Parameters
----------
spectrum_dict : Dict
The Pyteomics cluster dict to be parsed.
Returns
-------
MsmsSpectrum
The parsed cluster.
"""
spectrum_id = spectrum_dict['id']
mz_array = spectrum_dict['m/z array']
intensity_array = spectrum_dict['intensity array']
retention_time = spectrum_dict['scanList']['scan'][0]['scan start time']
precursor = spectrum_dict['precursorList']['precursor'][0]
precursor_ion = precursor['selectedIonList']['selectedIon'][0]
precursor_mz = precursor_ion['selected ion m/z']
if 'charge state' in precursor_ion:
precursor_charge = int(precursor_ion['charge state'])
elif 'possible charge state' in precursor_ion:
precursor_charge = int(precursor_ion['possible charge state'])
else:
raise ValueError('Unknown precursor charge')
return sus.MsmsSpectrum(spectrum_id, precursor_mz, precursor_charge,
mz_array, intensity_array, None, retention_time)
def _parse_ms2lda(usi: str) -> Tuple[sus.MsmsSpectrum, str]:
match = _match_usi(usi)
ms2lda_task_match = ms2lda_task_pattern.match(match.group(2))
if ms2lda_task_match is None:
raise UsiError("Incorrectly formatted MS2LDA task", 400)
experiment_id = ms2lda_task_match.group(1)
index_flag = match.group(3)
if index_flag.lower() != "accession":
raise UsiError("Currently supported MS2LDA index flags: accession",
400)
index = match.group(4)
try:
lookup_request = requests.get(
f"{MS2LDA_SERVER}get_doc/?experiment_id={experiment_id}"
f"&document_id={index}",
timeout=timeout,
)
lookup_request.raise_for_status()
spectrum_dict = json.loads(lookup_request.text)
if "error" in spectrum_dict:
raise UsiError(f'MS2LDA error: {spectrum_dict["error"]}', 404)
mz, intensity = zip(*spectrum_dict["peaks"])
source_link = f"http://ms2lda.org/basicviz/show_doc/{index}/"
spectrum = sus.MsmsSpectrum(usi, float(spectrum_dict["precursor_mz"]),
0, mz, intensity)
return spectrum, source_link
except requests.exceptions.HTTPError:
raise UsiError("Unknown MS2LDA USI", 404)
def _parse_spectrum(spectrum_dict: Dict) -> sus.MsmsSpectrum:
"""
Parse the Pyteomics cluster dict.
Parameters
----------
spectrum_dict : Dict
The Pyteomics cluster dict to be parsed.
Returns
-------
MsmsSpectrum
The parsed cluster.
"""
identifier = spectrum_dict['params']['title']
mz_array = spectrum_dict['m/z array']
intensity_array = spectrum_dict['intensity array']
retention_time = float(spectrum_dict['params']['rtinseconds'])
precursor_mz = float(spectrum_dict['params']['pepmass'][0])
if 'charge' in spectrum_dict['params']:
precursor_charge = int(spectrum_dict['params']['charge'][0])
else:
raise ValueError('Unknown precursor charge')
return sus.MsmsSpectrum(identifier, precursor_mz, precursor_charge,
mz_array, intensity_array, None, retention_time)
def _parse_massbank(usi: str) -> Tuple[sus.MsmsSpectrum, str]:
match = _match_usi(usi)
index_flag = match.group(3)
if index_flag.lower() != "accession":
raise UsiError("Currently supported MassBank index flags: accession",
400)
index = match.group(4)
try:
lookup_request = requests.get(f"{MASSBANK_SERVER}{index}",
timeout=timeout)
lookup_request.raise_for_status()
spectrum_dict = lookup_request.json()
mz, intensity = [], []
for peak in spectrum_dict["spectrum"].split():
peak_mz, peak_intensity = peak.split(":")
mz.append(float(peak_mz))
intensity.append(float(peak_intensity))
precursor_mz = 0
for metadata in spectrum_dict["metaData"]:
if metadata["name"] == "precursor m/z":
precursor_mz = float(metadata["value"])
break
source_link = (f"https://massbank.eu/MassBank/"
f"RecordDisplay.jsp?id={index}")
spectrum = sus.MsmsSpectrum(usi, precursor_mz, 0, mz, intensity)
return spectrum, source_link
except requests.exceptions.HTTPError:
raise UsiError("Unknown MassBank USI", 404)
def _parse_spectrum(spectrum_dict: Dict) -> sus.MsmsSpectrum:
"""
Parse the Pyteomics cluster dict.
Parameters
----------
spectrum_dict : Dict
The Pyteomics cluster dict to be parsed.
Returns
-------
MsmsSpectrum
The parsed cluster.
"""
spectrum_id = spectrum_dict['id']
mz_array = spectrum_dict['m/z array']
intensity_array = spectrum_dict['intensity array']
retention_time = spectrum_dict['retentionTime']
precursor_mz = spectrum_dict['precursorMz'][0]['precursorMz']
if 'precursorCharge' in spectrum_dict['precursorMz'][0]:
precursor_charge = spectrum_dict['precursorMz'][0]['precursorCharge']
else:
raise ValueError('Unknown precursor charge')
return sus.MsmsSpectrum(spectrum_id, precursor_mz, precursor_charge,
mz_array, intensity_array, None, retention_time)
def test_init_mz_sorted():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity)
for mz1, mz2 in zip(spec.mz[:-1], spec.mz[1:]):
assert mz1 <= mz2
def test_scale_intensity_rank_more_peaks():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity)
with pytest.raises(ValueError):
spec.scale_intensity(scaling='rank', max_rank=num_peaks - 50)
def test_annotate_peptide_fragments():
fragment_tol_mass = 0.02
fragment_tol_mode = 'Da'
peptides = [
'SYELPDGQVITIGNER', 'MFLSFPTTK', 'DLYANTVLSGGTTMYPGIADR', 'YLYEIAR',
'VAPEEHPVLLTEAPLNPK'
]
for peptide in peptides:
fragment_mz = np.asarray([
fragment.calc_mz for fragment in
spectrum._get_theoretical_peptide_fragments(peptide)
])
fragment_mz += np.random.uniform(-0.9 * fragment_tol_mass,
0.9 * fragment_tol_mass,
len(fragment_mz))
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
mz[:len(fragment_mz)] = fragment_mz
intensity = np.random.lognormal(0, 1, num_peaks)
charge = 2
spec = spectrum.MsmsSpectrum('test_spectrum',
mass.calculate_mass(sequence=peptide,
charge=charge),
charge,
mz,
intensity,
peptide=peptide)
spec.annotate_peptide_fragments(fragment_tol_mass, fragment_tol_mode)
assert np.count_nonzero(spec.annotation) == len(fragment_mz)
def _parse_gnps_task(usi: str) -> Tuple[sus.MsmsSpectrum, str]:
match = _match_usi(usi)
gnps_task_match = gnps_task_pattern.match(match.group(2))
if gnps_task_match is None:
raise UsiError("Incorrectly formatted GNPS task", 400)
task = gnps_task_match.group(1)
filename = gnps_task_match.group(2)
index_flag = match.group(3)
if index_flag.lower() != "scan":
raise UsiError("Currently supported GNPS TASK index flags: scan", 400)
scan = match.group(4)
try:
request_url = (f"https://gnps.ucsd.edu/ProteoSAFe/DownloadResultFile?"
f"task={task}&invoke=annotatedSpectrumImageText&block=0"
f"&file=FILE->{filename}&scan={scan}&peptide=*..*&"
f"force=false&_=1561457932129&format=JSON")
lookup_request = requests.get(request_url, timeout=timeout)
lookup_request.raise_for_status()
spectrum_dict = lookup_request.json()
mz, intensity = zip(*spectrum_dict["peaks"])
source_link = (f"https://gnps.ucsd.edu/ProteoSAFe/status.jsp?"
f"task={task}")
if "precursor" in spectrum_dict:
precursor_mz = float(spectrum_dict["precursor"].get("mz", 0))
charge = int(spectrum_dict["precursor"].get("charge", 0))
else:
precursor_mz, charge = 0, 0
spectrum = sus.MsmsSpectrum(usi, precursor_mz, charge, mz, intensity)
return spectrum, source_link
except (requests.exceptions.HTTPError, json.decoder.JSONDecodeError):
raise UsiError("Unknown GNPS task USI", 404)
def _read_spectra_mgf(filename: str) -> Iterable[sus.MsmsSpectrum]:
"""
Read MS/MS spectra from an MGF file.
Parameters
----------
filename : str
The MGF file name.
Returns
-------
Iterable[sus.MsmsSpectrum]
An iterable of spectra in the given MGF file.
"""
for spectrum_dict in tqdm.tqdm(pyteomics.mgf.read(filename),
desc='Spectra read', unit='spectra'):
spectrum = sus.MsmsSpectrum(
spectrum_dict['params']['title'],
spectrum_dict['params']['pepmass'][0],
spectrum_dict['params']['charge'][0],
spectrum_dict['m/z array'],
spectrum_dict['intensity array'],
None,
spectrum_dict['params'].get('rtinseconds'))
spectrum.filename = spectrum_dict['params'].get(
'filename', os.path.splitext(os.path.basename(filename))[0])
if 'scan' in spectrum_dict['params']:
spectrum.scan = spectrum_dict['params']['scan']
if 'cluster' in spectrum_dict['params']:
spectrum.cluster = spectrum_dict['params']['cluster']
yield spectrum
def plot_prediction(
self,
peptide,
modifications,
charge,
prediction=None,
ax=None,
filename=None,
):
"""
Plot MS²PIP-predicted spectrum with spectrum_utils.
Parameters
----------
peptide: string
Unmodified peptide sequence. Only canonical amino acids are allowed, and
peptide sequence should be of length [3, 100].
modifications: string
MS²PIP style-formatted modification string (e.g. `0|Acetyl|5|Oxidation`).
See MS²PIP README.md for more info.
charge: int
Peptide precursor charge.
prediction: tuple or None (default: None)
Tuple with `ms2pip.single_prediction.SinglePrediction.predict()` output.
ax: matplotlib.axes.Axes or None (default: None)
Figure ax to plot onto.
filename: str or None (default: None)
Filename to save plot to. File extension defines the format. Figure will
not be saved if None.
"""
if not prediction:
prediction = self.predict(peptide, modifications, charge)
mz, intensity, annotation = prediction
identifier = f"{peptide}/{charge}/{modifications}"
precursor_mz = self.mod_info.calc_precursor_mz(peptide, modifications,
charge)
mod_dict = self._modifications_to_dict(modifications)
sus_annotation = self._get_sus_annotation(mz, annotation)
spectrum = sus.MsmsSpectrum(
identifier,
precursor_mz,
charge,
mz,
intensity,
annotation=sus_annotation,
retention_time=None,
peptide=peptide,
modifications=mod_dict,
)
if not ax:
ax = plt.gca()
sup.spectrum(spectrum, ax=ax)
ax.set_title("MS²PIP prediction for " + identifier)
if filename:
plt.savefig(filename)
def test_scale_intensity_rank():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity)
spec.scale_intensity(scaling='rank')
np.testing.assert_allclose(np.sort(spec.intensity),
np.arange(1, num_peaks + 1))
def test_mz_annotation_len():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
annotation = [str(this_mz) for this_mz in mz[:100]]
with pytest.raises(ValueError):
spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity,
annotation)
def test_round_no_merge():
num_peaks = 150
mz = np.arange(1, num_peaks + 1) + np.random.uniform(-0.49, 0.5, num_peaks)
intensity = np.random.exponential(1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity)
spec.round(0)
for i, this_mz in enumerate(spec.mz, 1):
assert i == pytest.approx(this_mz)
def _parse_motifdb(usi: str) -> Tuple[sus.MsmsSpectrum, str]:
# E.g. mzspec:MOTIFDB:motif:motif_id.
tokens = usi.split(':')
motif_id = tokens[3]
request_url = f'{MOTIFDB_SERVER}get_motif/{motif_id}'
mz, intensity = zip(*json.loads(requests.get(request_url).text))
source_link = f'http://ms2lda.org/motifdb/motif/{motif_id}/'
return sus.MsmsSpectrum(usi, 0, 1, mz, intensity), source_link
def _dict_to_spectrum(spectrum_dict: Dict):
return sus.MsmsSpectrum(
spectrum_dict['params']['title'],
spectrum_dict['params']['pepmass'][0],
spectrum_dict['params']['charge'][0],
spectrum_dict['m/z array'],
spectrum_dict['intensity array'],
retention_time=spectrum_dict['params']['rtinseconds'])
def test_init_modification():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
modifications = {0: +1, 4: -56, 'N-term': +16, 'C-term': -42}
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity,
peptide='PEPTIDER',
modifications=modifications)
assert len(spec.modifications) == 4
def test_init_annotation_order():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
annotation = [str(this_mz) for this_mz in mz]
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity,
annotation)
for this_mz, this_annotation in zip(spec.mz, spec.annotation):
assert this_mz == pytest.approx(float(this_annotation))
def test_filter_intensity_keep_all():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity)
spec.filter_intensity()
assert len(spec.mz) == num_peaks
assert len(spec.intensity) == num_peaks
assert len(spec.annotation) == num_peaks
def test_annotate_molecule_fragments_invalid_mz():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 1, mz, intensity)
fragment_tol_mass = 0.02
fragment_tol_mode = 'Da'
with pytest.raises(ValueError):
spec.annotate_molecule_fragment('CCCCCCCC', 1600, 1, fragment_tol_mass,
fragment_tol_mode)
def test_round_merge_max():
num_peaks = 10
mz = np.arange(1, num_peaks + 1) + np.random.uniform(-0.2, 0.2, num_peaks)
mz[4] = mz[3] + 0.0002
mz[5] = mz[3] + 0.0005
mz[7] = mz[8] - 0.00037
intensity = np.arange(1, 11)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity.copy())
spec.round(1, 'max')
np.testing.assert_allclose(spec.intensity, [1, 2, 3, 6, 7, 9, 10])
def test_round_merge_sum():
num_peaks = 10
mz = np.arange(1, num_peaks + 1) + np.random.uniform(-0.2, 0.2, num_peaks)
mz[4] = mz[3] + 0.0002
mz[5] = mz[3] + 0.0005
mz[7] = mz[8] - 0.00037
intensity = np.random.exponential(1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity.copy())
spec.round(1, 'sum')
assert np.sum(spec.intensity) == pytest.approx(np.sum(intensity))
def test_init_intensity_order():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
mz_intensity_tuples = sorted(zip(mz, intensity),
key=operator.itemgetter(0))
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity)
for this_mz, this_intensity, mz_intensity_tuple in zip(
spec.mz, spec.intensity, mz_intensity_tuples):
assert (this_mz, this_intensity) == pytest.approx(mz_intensity_tuple)
def test_scale_intensity_max():
num_peaks = 150
mz = np.random.uniform(100, 1400, num_peaks)
intensity = np.random.lognormal(0, 1, num_peaks)
spec = spectrum.MsmsSpectrum('test_spectrum', 500, 2, mz, intensity)
intensity_copy, max_intensity = spec.intensity.copy(), spec.intensity.max()
spec.scale_intensity(max_intensity=1.)
assert spec.intensity.max() == pytest.approx(1.)
np.testing.assert_allclose(spec.intensity * max_intensity, intensity_copy,
rtol=1e-5)